Year summary - 2023

Posted by Tobiasz Kedzierski on 01.01.2024

2023

Highlights

git rebase --interactive and fixups

Presented a well-received talk at work on mastering git rebase --interactive with a focus on fixups. Inspired by Raymond Hettinger I decided to use the documentation generator MkDocs and Material for MkDocs. The presentation is available as a website under my domain: tobked.dev/git_rebase_interactive. Its source is on my Github: github.com/TobKed/git_rebase_interactive.

git rebase

Google Cloud Developer Day Warsaw 2023

I attended Google Cloud Developer Day Warsaw 2023 on April 27th, 2023. It was a free event organized by Google Cloud at their Warsaw Campus. I wrote more about it in a separate blog post: Google Cloud Developer Day Warsaw 2023.

GitOps and Flux

I dived into the world of flux, a game-changer for GitOps deployments. This powerful tool automates keeping your Kubernetes cluster in sync with your git configuration. It goes beyond just watching for changes in configuration files - Flux can also identify new Docker images and update both configurations and the cluster itself. Learning Flux has been a blast, and it's a valuable addition to my DevOps toolkit.

Taking ML Pipelines to Production: Journey with Kubeflow and Vertex Pipelines

I embarked on a journey to automate and streamline my machine learning workflows. Kubeflow Pipelines, a powerful open-source platform allows for building and managing complex ML pipelines, encompassing everything from data ingestion to model training and deployment.

The real magic happened when I integrated Kubeflow Pipelines with Vertex Pipelines on Google Cloud. Vertex Pipelines is a managed service that makes running Kubeflow Pipelines in a production environment a breeze. This powerful combination allowed me to successfully deploy and run ML pipelines in production, ensuring scalability and reliability.

By leveraging Kubeflow Pipelines and Vertex Pipelines, I've gained the ability to automate ML workflows, improve their reproducibility, and confidently run them in a production setting. This has significantly boosted my machine learning development efficiency.

Google Cloud and Pycon CZ 2023 presentation

One of the highlights was diving deep into Google Cloud Platform (GCP) to build a highly scalable application. CloudRun streamlined code deployment, while Cloud SQL (PostgreSQL) provided a robust and scalable foundation for my app's data. PubSub messaging became a game-changer, decoupling APIs for independent scaling and boosting overall performance. A load balancer added the finishing touch, ensuring smooth traffic distribution and regional optimization. Seeing the app handle traffic peaks with minimal latency was incredibly rewarding.

I had the pleasure of attending the incredible PyCon CZ conference in Prague. The excitement began even before I arrived in Prague, as I had the chance to meet John and Judy from the USA during my train journey. John, a retired programmer, shared stories of his work with natural language, SQL-Windows, and COBOL. It felt like I was meeting a living legend!

During the train ride, I also had the pleasure of getting to know Piotr, the organizer of PyKonik. The conference itself was a fantastic experience, with a multitude of great people and captivating presentations.

One of the best moments of the conference for me was when I put myself on the list for the Lightning Talks. This was a huge challenge for me as I am usually quite nervous when it comes to public speaking, but I was determined to push myself out of my comfort zone. The title of my presentation was "ML (Fast)API below 100ms (on GCP)". The presentation, as well as all the resources related to it, can be found in my repository. I have provided an end-to-end working example on Google Cloud that demonstrates how to defer load from FastAPI (served by CloudRun) to a background worker (also on CloudRun) via a queueing system (PubSub). After practicing my presentation numerous times in my hotel room, I finally stepped onto the stage and gave it my all. It was a personal victory over my fear of public speaking.

PyCon CZ.

Python API and background workers on Google Cloud Platform repo

Writing

I've been writing monthly summaries, compiling a few sentences about the past month alongside articles, tools, and other interesting finds.

Each month, I add a personal touch with a photo I've taken

2023 summaries:

I also wrote one additional blog post:

Summary

This year has been a fantastic journey of learning, pushing boundaries, and contributing to the programming community. I'm excited to see what the next year holds!

Photos

2023_01

2023_04

2023_05

2023_06

2023_07

2023_08

2023_09

2023_10

2023_11

2023_12

tags: summary, year, 2023